Abstract

The present study deals with Landsat data for the change analysis and flood mapping using water class extraction. Among the various available bands, the most suitable band combination is identified using class separability analysis and a novel water index (WI₂₀₂₀) is proposed for the extraction of water. The proposed index is able to separate the water class from other land features and performs better than NDWI. The index-based results are also validated through supervised classification and used for change analysis and flood mapping. The change analysis is done in two ways, viz., from the year 2000 to 2020 in a span of 5 years in a periodic manner and from the year 2014 to 2020 in a continuous manner. The flood mapping is done on the images of 2014 and 2016 along with the pre-flood and post-flood situations with the help of WI₂₀₂₀ which gives quantitatively satisfactory results.

摘要

本研究使用水类提取处理用于变化分析和洪水制图的 Landsat 数据。在各种可用波段中，使用类别可分离性分析和新的水指数 (WI ₂₀₂₀ ) 确定了最合适的波段组合) 建议用于提取水。建议的指数能够将水类与其他土地特征分开，并且比 NDWI 表现更好。基于指数的结果还通过监督分类进行验证，并用于变化分析和洪水测绘。变化分析以两种方式进行，即从2000年到2020年以5年为周期周期性进行，从2014年到2020年以连续方式进行。在 WI ₂₀₂₀的帮助下，对 2014 年和 2016 年的图像以及洪水前和洪水后的情况进行了洪水测绘，结果在数量上令人满意。